# LedgerProof LAIN — Verified Evidence for AI Agents (MCP)

**LAIN** (LedgerProof Autonomous Intelligence Network) is a remote MCP server that sells
**cryptographically anchored evidence**: verified agent-run receipts and existence proofs.
Every deliverable's hash is anchored in a public, append-only log and verifies
independently of LedgerProof's availability.

The settlement rule is enforced in code: **payment captures only after the anchored
receipt exists.** Failed or unproven work is never charged.

## Connect

```json
{
  "type": "streamable-http",
  "url": "https://ledgerproofhq.io/api/lain/mcp"
}
```

Listed in the official MCP Registry as [`io.ledgerproofhq/lain`](https://registry.modelcontextprotocol.io/v0/servers?search=ledgerproof).

## Tools

| Tool | What it does |
|---|---|
| `discover_capabilities` | Capabilities, input schemas, pricing, environments, settlement rule |
| `request_capability` | Buy a capability — e.g. `receipt.agent-run`: pass `action` + one of `inputs_hash`/`outputs_hash`/`state_hash` (sha256) |
| `get_job` | Poll a job; for paid checkout jobs this triggers proof-gated capture and returns the deliverable |

## Environments

- **`test`** (default) — simulated money; try the full flow free, including machine-to-machine settlement.
- **`live`** — a real purchase: the response returns the anchored proof immediately plus a
  Stripe checkout URL; the deliverable releases when payment captures against the proof.

## What a receipt attests

Existence (this run metadata existed at `observed_at`) and integrity (any later
alteration breaks verification). It deliberately does **not** attest that the run
occurred as described, nor the quality of the agent's work — that scope is stated
machine-readably inside every receipt.

## Discovery

- HTTP API: `https://ledgerproofhq.io/api/lain/capabilities`
- Well-known manifest: `https://ledgerproofhq.io/.well-known/capabilities.json`

---

© LedgerProof · [ledgerproofhq.io](https://ledgerproofhq.io) · verifying is free, forever.